In Camdeboo National Park (South Africa), Using Fixed-Point Photo Monitoring: the Role of Herbivory and Climate

Page 1 of 16 Original Research

Vegetation change (1988–2010) in Camdeboo National Park (South Africa), using fixed-point photo monitoring: The role of herbivory and climate

Authors: Fixed-point photo monitoring supplemented by animal census data and climate monitoring 1 Mmoto L. Masubelele potential has never been explored as a long-term monitoring tool for studying vegetation Michael T. Hoffman1 William Bond1 change in the arid and semi-arid national parks of South Africa. The long-term (1988–2010), Peter Burdett2 fixed-point monitoring dataset developed for the Camdeboo National Park, therefore, provides an important opportunity to do this. Using a quantitative estimate of the change in Affiliations: 1Department Botany, vegetation and growth form cover in 1152 fixed-point photographs, as well as series of step- University of Cape Town, point vegetation surveys at each photo monitoring site, this study documented the extent of South Africa vegetation change in the park in response to key climate drivers, such as rainfall, as well as 2Camdeboo National Park, land use drivers such as herbivory by indigenous ungulates. We demonstrated the varied Graaff-Reinet, South Africa response of vegetation cover within three main growth forms (grasses, dwarf shrubs [< 1 m] and tall shrubs [> 1 m]) in three different vegetation units and landforms (slopes, plains, Correspondence to: Mmoto Masubelele rivers) within the Camdeboo National Park since 1988. Sites within Albany Thicket and Dwarf Shrublands showed the least change in vegetation cover, whilst Azonal vegetation Email: and Grassy Dwarf Shrublands were more dynamic. Abiotic factors such as drought and mmotomasubelele@gmail. com flooding, total annual rainfall and rainfall seasonality appeared to have the greatest influence on growth form cover as assessed from the fixed-point photographs. Herbivory appeared Postal address: not to have had a noticeable impact on the vegetation of the Camdeboo National Park as far PO Box 216, Steenberg 7947, as could be determined from the rather coarse approach used in this analysis and herbivore South Africa densities remained relatively low over the study duration. Dates: Received: 19 Nov. 2012 Conservation implications: We provided an historical assessment of the pattern of vegetation Accepted: 05 Aug. 2013 and climatic trends that can help evaluate many of South African National Parks’ biodiversity Published: 18 Oct. 2013 monitoring programmes, especially relating to habitat change. It will help arid parks in How to cite this article: assessing the trajectories of vegetation in response to herbivory, climate and management Masubelele, M.L., Hoffman, interventions. M.T., Bond, W. & Burdett, P., 2013, ‘Vegetation change (1988–2010) in Camdeboo National Park Introduction (South Africa), using fixed- point photo monitoring: Preserving and conserving natural areas and reducing biodiversity loss are core principles of most The role of herbivory and protected area management programmes (Bruner et al. 2001; Ehrlich & Pringle 2008; Rodrigues climate’, Koedoe 55(1), Art. et al. 2004). Protected areas managers across the world are confronted with a range of critical #1127, 16 pages. http:// dx.doi.org/10.4102/koedoe. decisions in terms of assessing and anticipating climate change impacts as well as land use impacts v55i1.1127 on biodiversity (Maraiano, Falcucci & Boitani 2008). These management decisions and solutions, however, are best made if they are informed by reliable monitoring data which document the Note: extent, nature and rate of environmental change over time as well as the key potential drivers of Additional supporting information may be found such change (Gitzen et al. 2012; Lindenmayer & Likens 2010; Scholes et al. 2008). in the online version of this article as an Online The Camdeboo National Park (CNP), which surrounds the town of Graaff-Reinet in the semi-arid Appendix: http://dx.doi. Karoo Midlands, South Africa, achieved national park status in 2006 and conserves remnants of org/10.4102/koedoe. v55i1.1127-1. the main biomes in the region. For 27 years prior to this, the area was managed as a provincial nature reserve after it had been used as a communal grazing area for several decades (Burdett Copyright: 1995). The impact of previous land uses has left a legacy of erosion and degradation in some parts © 2013. The Authors. of the landscape and active attempts have been made to restore these areas (CNP Newsletter Licensee: AOSIS OpenJournals. This work 2011). Several studies document the natural resources of the CNP, particularly the vegetation is licensed under the (Palmer 1990). However, these studies were undertaken decades ago and critical issues such as Creative Commons the recent rate of spread and impact of alien plants within different vegetation types in the CNP Attribution License. is not known (CNP Management Plan 2006; Masubelele, Foxcroft & Milton 2009). Read online: Scan this QR Outside the park, within the commercial livestock producing areas, the influence of rainfall and code with your smart phone or grazing by domestic animals has been well documented (Hoffman, Barr & Cowling 1990; O’Connor mobile device & Roux 1995). However, few studies have been conducted on the impact of wild herbivores on to read online. rangeland condition within either private or state-owned protected areas in the region. In the

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 2 of 16 Original Research

Karoo National Park, about 150 km west of the CNP, a change Repeat fixed-point photography offers an important tool in land use from small-stock farming to conservation and for monitoring vegetation change, particularly because the reintroduction of wild ungulates resulted in an increase the spatial scale of analysis ranges from local to landscape in perennial grass cover following a change in land use. scale and is of direct relevance to park managers (Hall 2000; However, rates of recovery were slower in areas exposed to Nichols, Ruyle & Nourbakhsh 2009; Turner et al. 2003). The heavy indigenous ungulate herbivory (Kraaij & Milton 2006). potential of fixed-point photo monitoring has also never Whilst a census of indigenous ungulates within the CNP has been explored as a long-term monitoring tool for use in the been conducted since 1990 there has not been any synthesis of arid and semi-arid national parks of South Africa. Owing to these data. Little is known, therefore, of the concentration and the diversity of vegetation types in the CNP, it provides a impact of indigenous ungulates within different vegetation useful case study to explore the potential of this technique types of the park. Without this information it is difficult to for monitoring vegetation change. A repeat fixed-point photography monitoring programme was established in determine the impact of indigenous ungulate herbivory on the CNP in 1988 but the efficacy and value of the results the vegetation of the reserve. have not been assessed. This study is important because if it proves successful it will be recommended to SANParks Another major concern for the conservation of biodiversity as an appropriate monitoring tool for the assessment and in protected areas is the impact that future climate change monitoring of environmental change over long time frames. will have on biodiversity and ecosystem function (Perrings et al. 2011). For example, rainfall is the key abiotic driver in The main objective of the study was to document the the Karoo Midlands (O’Connor & Roux 1995) and extended extent of vegetation change in the CNP in response to key droughts are common in the region. It is important for climate drivers, such as rainfall, as well as land use drivers, managers to understand the influence of periodic floods especially herbivory by indigenous ungulates. A secondary and droughts on vegetation structure and composition over objective was to evaluate the use of fixed-point photography time frames of several years and decades. An understanding as a monitoring tool for studying change in the main growth of current rainfall–vegetation dynamics will enable forms and vegetation types at the CNP. In this synthesis of conservation managers to plan more effectively for future 22 years of fixed-point monitoring data, we have addressed changes in climate. Climate change projections for the the following main questions, (1) ‘How has growth form region suggest that the area will become more arid with an composition changed within different vegetation types and increase in the occurrence of more extreme events such as land forms?’, (2) ‘What influence have drought and flooding drought (Department of Environmental Affairs 2010; Driver events as well as changes in seasonal rainfall patterns had et al. 2012; Midgley et al. 2002). Thicket vegetation around on these patterns?’, (3) ‘How has herbivore density and Graaff-Reinet, for example, is expected to be amongst the concentration influenced vegetation cover and growth worst affected of all Albany thicket types in the Eastern Cape form composition?’ and (4) ‘Is fixed-point photography an (Robertson & Palmer 2002). Current projections also suggest appropriate monitoring tool for addressing key management an increase in shrub cover at the expense of grasses, with concerns within SANParks’ arid and semi-arid conservation obvious knock-on effects for vegetation communities of the areas?’. Nama-karoo biome, as well as for grazer–browser herbivore ratios in the CNP (Midgley et al. 2002; Rutherford, Powrie & Research method and design Husted 2012a, 2012b). Study area The CNP surrounds the town of Graaff-Reinet (Figure 1), Monitoring is an important tool for management as it which is situated in the Camdeboo Municipality of the helps in understanding how ecosystems are structured Eastern Cape Province of South Africa. The CNP is located and how they function over time (Noss 1990). There has between 32º18’14.83”S and 24º38’31.41”E and lies from been a surge in development of long-term biodiversity 740 m.a.s.l. to 1480 m.a.s.l. at the foothills of the Sneeuberg monitoring programmes to address some of the millennium range, with a small section of low-lying plains included development goals and also to respond to climate change for within the park boundaries. The climate is semi-arid with most countries in the world (Scholes et al. 2008). According to 32.0% of the average annual total of 336 mm of rain Reyers and McGeoch (2007), southern Africa has responded (CV = 31.5%) falling during the hottest months of the year with the initiation of the South African Earth Observation (February–April). The CNP also experiences snow, fog Network (SAEON), as well as Biodiversity monitoring and frost, with maximum temperatures during summer of transect analysis (BIOTA). Recently, South African National 43 ºC and minimum temperatures of -3 ºC in winter (CNP Parks (SANParks) has established a biodiversity monitoring Management Plan 2006). The mean annual maximum framework (McGeoch et al. 2011), however, few national temperature during the study period was 26 ºC, whereas mean parks, not only in South Africa but indeed throughout annual minimum temperature was 10 ºC. The hydrology of the world, have a comprehensive, long-term monitoring the CNP is largely based on its location at the edge of the programme in place, especially for the vegetation. Whilst the Great Escarpment, from which six seasonal rivers (Sundays, Kruger National Park burning plots represent an exception Gats, Melk, Camdeboo, Pienaars and Erasmuskloof) drain (Trollope et al. 1998), most studies on vegetation change in into the Nqweba Dam in the central area of the park (CNP protected areas are of less than 20 years duration. Management Plan 2006).

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 3 of 16 Original Research

Monitoring sites Rainfall stations Camdeboo new section Nqweba dam N9 Highway

W E

S 0 1 2 4 6 8 Kilometres

FIGURE 1: Location and extent of the Camdeboo National Park at Graaff-Reinet in the Eastern Cape Province of South Africa.

The geological system of the CNP consists of very thick Shrublands comprise Nama-karoo biome vegetation types. layers of near-horizontal strata of sedimentary rocks of Albany Thicket, which was called Succulent Thicket by tertiary origin. Large areas of the park, however, are covered Palmer (1990), represents vegetation associated with the by alluvium, gravel, sand, mud and stone of more recent Albany Thicket biome. In addition to the three physiognomic origin. Soils derived from these quaternary deposits have an classes outlined above, a fourth Azonal unit located within important influence on the vegetation in the park because the ephemeral rivers and Nqweba Dam flood plain was also they represent the growth medium for many dwarf shrubs recognised. in the region (Lovegrove 1993). The soils are generally calcareous duplex forms of secondary nature having been Although the CNP was first proclaimed and managed as deposited as alluvium on the impermeable sandstone. They a provincial reserve (called the Karoo Nature Reserve) are subject to sheet and gully erosion, which is aggravated in 1979, it became a national park in 2006 and is the 22nd when vegetation cover is reduced. protected area to fall under the management of SANParks. The area has a very long history of land use. Immediately The vegetation of the CNP falls into three biomes, namely prior to the proclamation of the Karoo Nature Reserve in the Albany Thicket, Grassland and Nama-karoo biomes 1979, the land was leased to tenant farmers who grazed their livestock on the property, thus contributing to overgrazing (Mucina & Rutherford 2006). Although Palmer (1990) and the erosion of some areas (Burdett 1995; CNP provided a detailed phytosociological analysis of the former Management Plan 2006). Karoo Nature Reserve, the main vegetation types currently recognised in the CNP follow Mucina and Rutherford’s (2006) classification. These are: Camdeboo Escarpment Photo monitoring procedure Thicket (AT14), Karoo Escarpment Grassland (Gh1), Eastern The fixed-point photo monitoring system used in this study Lower Karoo (NKI2), Upper Karoo Hardeveld (NKu2) and was started in 1988 by the two nature reserve managers at the Eastern Upper Karoo (Nku4). Azonal vegetation also occurs time (Ken Coetzee and Peter Burdett) with the establishment around the Nqweba Dam, which is classified by Mucina of initial 15 sites (Table 1). Additional sites were added in and Rutherford (2006) as Southern Karoo Riviere (AZi6). 1989 and 1992 by the current park manager (Peter Burdett) In this study, the vegetation types outlined above have within each of the four main physiognomic classes, to reach been grouped into three broad physiognomic classes for a total of 32 sites. The main objective of the monitoring analysis. These classes are broadly reflective of the major programme was to document the impact of herbivores on biomes in the region. Grassy Dwarf Shrublands represent the different vegetation types of the reserve. Between 1988 vegetation types within the Grassland biome, whilst Dwarf and 2005 each site was photographed a total of 12 times.

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 4 of 16 Original Research

After the proclamation of the CNP in 2006, the sites were Vegetation classification and the grouping of not re-photographed until 2010 when this current study was sample sites undertaken. This was largely because of the time and effort Vegetation cover data were used to support the subjective required by the park manager to effect the transition of the grouping of sites into four vegetation units by the park Karoo Nature Reserve into the CNP and to align the newly manager. We used species composition and cover data acquired conservation area with broader SANPark’s policies. collected using a step-point sampling technique during 2010. The step-point surveys were carried out along a 100 m Each of the 32 sites was permanently marked with a 70 cm transect along each direction (north, east, south, west) at a square concrete block set in the ground to a depth of about photo location. To show the grouping of these 128 transects 50 cm. In the centre of the concrete block, a four-sided (32 sites × four directions), a non-metric multidimensional metal pole was fixed and allowed to protrude about 40 cm scaling (NMS) ordination procedure using a Sorensen above the surface in such a way that each of the sides of the distance measure was undertaken within PC-Ord version metal pole faced in one of four directions (north, east, south 5.33 (McCune & Mefford 2006). This ordination technique and west). This permanently fixed basal pole was used to is suitable for the representation of ecological data because position a removable 170 cm high metal pole with a slightly it makes no assumptions of linearity or underlying models larger diameter which fitted securely over the basal pole. A of species relationships. NMS iteratively searches for flat plate, used to secure the camera was welded on the top of ordering or grouping of samples in multidimensional space the removable pole. that minimises the stress and number of dimensions in the solution. The stress value provides a measure of the goodness of fit of the data to the ordination solution and follows A standard 50 mm lens mounted on a manual single lens Kruskal’s (1964) formula as rescaled by Mather (1976). Each reflex (SLR) camera loaded with 35 mm colour slide film of the 128 transects was also grouped subjectively into one was used for most of the years prior to 2010, except in of three landforms in which they occurred: rivers, plains and 2003 when a wider-angle (35 mm) lens was used at some slopes. Change in the cover of different growth forms within of the sites. To replicate this process in 2010, a 50 mm lens, each of these landforms was assessed over the sampling mounted on a Minolta X300s manual SLR camera and loaded period. with Fujichrome Sensia 35 mm colour slide film, was used. However, because of the ease of postproduction processing Climate analysis and the saving in cost, a digital camera was also used (Canon Powershot Gll). All the original and 2010 colour slide images Annual (October–September) and monthly rainfall patterns were scanned as TIFF images at 300 dpi, A3 output format were analysed from data collected at four weather stations using a high quality Nikon Coolscan 5000 scanner at the Plant established within different sections of the CNP since 1987 by Conservation Unit, University of Cape Town. This amounted the park manager (Figure 1). Change in annual rainfall was assessed using a Mann Kendall non-parametric test for trend to a total of 1152 images and an electronic copy of all of the in time series data (Modarres & Da Silva 2007). A standardised photographs together with the metadata was sent to the CNP precipitation index (SPI) was also calculated to determine manager to whom the original material was also returned. the change over time in the incidence of drought and wet periods (Mckee, Doesken & Kleist 1993). To assess the extent Repeat photos were matched as accurately as possible and of change in rainfall seasonality, monthly rainfall data were analysed in Photoshop CS4. The cover of grasses, dwarf grouped into two time periods (1990–2000 and 2000–2010). shrubs (< 1 m) and tall shrubs (> 1 m) were estimated using In addition, the average values for the months of December, expert analysis for each individual image. Three ecologists, January and February for the four stations were summed with field knowledge of the vegetation of the region, to represent early summer rainfall. These values were then independently estimated the cover for each growth form on compared for each decade and tested for significance using each of the images obtained in the different vegetation units. a univariate analysis of variance. This analysis was repeated This was performed separately for the north, east, south and for autumn (March–May), winter (June–August) and spring west directions and then averaged to provide an individual (September–November) time periods. Long-term A-Pan site value for each growth form. Values for similar vegetation evaporation data obtained from the Department of Water units were then grouped together to get the average value for Affairs’ website for the period (1989–2008) for Nqweba Dam, the vegetation unit at each time period. as well as the average maximum and minimum temperature

TABLE 1: The total number of sites in each vegetation unit and the year in which sites were established by the park manager at the Camdeboo National Park. Date that sites were Vegetation unit Total number of sites established Grassy Dwarf Shrublands Dwarf Shrublands Albany Thicket Azonal 1988 4 3 5 3 15 1989 0 1 4 0 5 1992 3 4 1 4 12 Total number of sites 7 8 10 7 32

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 5 of 16 Original Research for Graaff-Reinet obtained from the South African Weather which are both located in the floodplain of the Nqweba Dam. Service (1992–2008) were also analysed. Changes in the New saplings, particularly of Acacia karroo, emerged at both temperature and A-Pan evaporation time series data were of these sites after 2004 and by 2010 had reached 2 m – 3 m in assessed using a Mann Kendall non-parametric test for trend Camdeboo National Park (Modarres & Da Silva 2007). Camdeboo National Park Vegetation type Axis2 Camdeboo National Park Vegetation type Azonal Grassy DwarfDwarf SchrublandsShrublands Axis2

Axis 2 Vegetation type Dwarf Shrublands Shrublands Azonal Herbivore abundance 1.0 Albany Thicket GrassyAlbany Dwarf Thicket Shrublands Dwarf Shrublands The location and abundance of different ungulate species 1.0 Albany Thicket were recorded annually by CNP field rangers from 1990 to 2010. Each species was converted to a large stock unit equivalent value (Furstenburg 2002; Meissner et al. 1983) and Axis 1 multiplied by the number of individuals recorded for each 0.0 species during each census event. During the annual census -1.0 0.0 1.0 2.0 AxisAxis 1 0.0 the location of each species was undertaken in the field using -1.0 0.0 1.0 2.0 a grid reference system created by the park manager. The GPS co-ordinates of the centre of each grid were subsequently determined from Google Earth. In this way each observation record was assigned a GPS coordinate. These GPS points -1.0 were used to create a shapefile in DIVA-GIS which was -1.0 then geo-referenced in ArcToolbox. Stocking density maps were then created for each annual census period in ArcGIS 9.3 (2008). The inverse distance weighted interpolation technique, which is a spatial analysis tool, was then used -2.0 to create a grazing impact surface for the CNP (Murwira & Skidmore 2005). In this technique the assumption is that -2.0 herbivory at a certain location also influences the vegetation in the neighbouring locations, because proximity of palatable N = 128 transects and 179 species. plants can increase the impact of herbivores on palatable and FIGURE 2: A non-metric multidimensional scaling ordination of the vegetation units at Camdeboo National Park, based on species cover data of the survey unpalatable plants in the surrounding area (Baraza, Zamora carried out in 2010. & Hòdar 2006). Average stocking density values for each of the 32 vegetation monitoring sites were also calculated to a determine how herbivore densities might have influenced 100 90 the vegetation of the CNP over time. 80 VM01 70 VM02 60 VM03 50 Results 40 VM23 30 VM26 Grass cover Grass 20 VM27

Grouping of sample sites % 10 VM28 0 Average Ordination results of the 2010 survey data from each 1992 1994 1995 1996 1998 2003 2004 2005 2010 of the four survey directions at the 32 photo locations Year grouped the sites into distinct vegetation units (Figure 2). A 100 b two-dimensional solution with the final stress value of 19.1 90 and a final instability value of 0.00025 was recommended 80 VM01 70 VM02 after 500 iterations. There is a clear separation of Azonal sites 60 VM03 50 from the other vegetation units. The analysis shows some VM23 40 VM26 support for the more subjective grouping of photo location 30 20 VM27 sites by the park manager into Grassy Dwarf Shrublands, 10 VM28 Dwarf Shrublands, Albany Thicket and Azonal vegetation shrubs of dwarf Cover 0 Average % 1992 1994 1995 1996 1998 2003 2004 2005 2010 units. Azonal sites were also the most diverse and spread out Year in the ordination diagram. Grassy Dwarf Shrublands were separated in ordination space from Albany Thicket. Dwarf 100 c 90 Shrublands sites were interspersed with both Grassy Dwarf 80 VM01 VM02 Shrubland and Albany Thicket sites. shrubs 70 60 VM03 tall 50 VM23 Change in growth form cover within different 40 VM26 30 VM27 vegetation units 20 Cover of Cover 10 VM28 Azonal vegetation % 0 Average Sites within Azonal vegetation experienced a decrease in 1992 1994 1995 1996 1998 2003 2004 2005 2010 Year tall shrub cover between 1998 and 2003 (Figure 3a–c). This FIGURE 3: Percentage cover per site (± s.d.) of, (a) grasses, (b) dwarf shrubs and pattern, however, is strongly influenced by what happened (c) tall shrubs in Azonal Vegetation as determined from an analysis of fixed-point at site VM03 (Figure 4) and, to a lesser extent, at site VM01, photo monitoring sites between 1992 and 2010.

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 6 of 16 Original Research

a b c

VM03 N 1992 Match VM03 N 1994 Match VM03 N 1996 Match d e f

VM03 N 1998Source Match: Photograph by Peter Burnett VM03 N 2003 Match FIGURE 4: Loss of tall shrubs (primarily Acacia karroo) at VM03 N showing the decline between 1998VM03 and N 20032010 Match and the subsequent increase in 2010 in the Azonal vegetation at the Camdeboo National Park. (a) 1992, (b) 1994, (c) 1996, (d) 1998, (e) 2003 and (f) 2010.

height. Tall shrub cover at the remainder of the sites within a the Azonal vegetation unit remained relatively unchanged 100 90 VM07 over the study period. Common tall shrub species within this 80 VM08 70 VM09 vegetation unit are A. karroo, Tamarix ramosissima and Salix 60 VM11 50 VM12 babylonica. Grass cover generally increased from 1996 to 2010, 40 VM16

with a slight decline in 2005 (Figure 3a–c). Different sites cover Grass 30 VM17 20 VM19 % varied both in terms of their cover of grasses as well as how 10 VM20 0 VM29 this changed over time. The increase in grass cover over the 1989 1992 1994 1995 1996 1998 2003 2004 2005 2010 Average study period was not mirrored by a decline in dwarf shrubs Year which remained relatively stable over time, except perhaps 100 b for site VM23. Much of the dwarf shrub cover comprised 90 VM07 80 alien shrubs such as Salsola tragus, Xanthium spinosum and 70 VM08 60 VM09 Xanthium strumarium. VM11 50 VM12 40 VM16 30 VM17 Albany Thicket 20 VM19 10 VM20 0 VM29 Tall shrub cover within Albany Thicket vegetation remained shrubs of dwarf Cover 1989 1992 1994 1995 1996 1998 2003 2004 2005 2010 Average % unchanged over the study period (Figure 5a–c and Figure 6). Year Common tall shrub species at these sites include Buddleja c saligna, Carissa haematocarpa, Diospyros austro-africana, Euclea 100 crispa, Euclea undulata, Grewia occidentalis, Grewia robusta, 90 VM07 80 VM08 Gymnosporia heterophylla, Portulacaria afra, Pappea capensis, 70 VM09 shrubs 60 VM11

Searsia longispina and Searsia undulata. The cover of dwarf tall 50 VM12 40 VM16 shrubs varied greatly between different sites within this 30 VM17 vegetation unit but remained relatively stable within a 20 VM19 10 VM20

Cover of Cover VM29 site over the study period. Common dwarf shrubs in this 0 Average % vegetation type include Becium burchellianum, Cheilanthes 1989 1992 1994 1995 1996 1998 2003 2004 2005 2010 Year

eckloniana, Crassula spp., Euryops spatheceus, Felicia filifolia and FIGUREFIGURE 5: 5: Percentage Percentage cover cover per per site site (± (± s.d.) s.d.) of, of, (a (a) grasses,) grasses, (b) (b) dwarf dwarf shrubs shrubs and and (c) (c) tall tall shrubs shrubs in inAlbany Albany Thicket Thicket FIGURE 5: Percentage cover per site (± s.d.) of, (a) grasses, (b) dwarf shrubs and Rhigozum obovatum. Although grass cover fluctuated during asFIGUREas determined determined 5: Percentage from from an an coveranalysis analysis per of siteof fixed-point fixed-point (± s.d.) of, photo photo(a) grasses, monitoring monitoring (b) dwarf sites sites betweenshrubs between and 1989 1989(c) talland and shrubs 2010. 2010. in Albany Thicket (c)as determined tall shrubs from in anAlbany analysis Thicket of fixed-point as determined photo monitoring from sites an between analysis 1989 of and fixed-point 2010. the course of the study, Albany Thicket vegetation was the photo monitoring sites between 1989 and 2010. FIGURE 5: Percentage cover per site (± s.d.) of, (a) grasses, (b) dwarf shrubs and (c) tall shrubs in Albany Thicket as determined from an analysis of fixed-point photo monitoring sites between 1989 and 2010.  http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127

FIGURE 5: Percentage cover per site (± s.d.) of, (a) grasses, (b) dwarf shrubs and (c) tall shrubs in Albany Thicket as determined from an analysis of fixed-point photo monitoring sites between 1989 and 2010. 

a b c

VM12 E 1993 Match VM12 E 1994 Match VM12 E 1996 Match d e f

VM12 E 1998Source Match : Photograph by Peter Burnett VM12 E 2010 Match VM12 E 2003 Match

AOSIS OpenJournals – Language Edited Version 14 Aug 2013 FIGURE 6: Albany Thicket example as shown by VM12 E depicting stability of the various growth forms between 1992 and 2010. (a) 1992,AOSISAOSIS OpenJournals OpenJournals (b) 1994, – – Language Language (c) 1996, Edited Edited Version Version(d) 1998,14 14 (e)Aug Aug 2013 2013 2003 and (f) 2010. FigureFigure 9 9 most stable vegetation type within the CNP, with very little a 100 change recorded in the cover of dwarf and tall shrubs. 90 80 VM04 70 VM05 Grassy Dwarf Shrublands and Dwarf Shrublands 60 VM13 50 For sites within both the Grassy Dwarf Shrubland (Figure 7a–c 40 VM14 30 VM22 and Figure 8) and Dwarf Shrubland (Figure 9a–c and Figure 10), cover Grass 20 VM24 periods of grass cover increase (e.g. between 1994 and 1997) % 10 VM25 0 were accompanied by a decline in the cover of dwarf shrubs 1992 1994 1995 1996 1997 1998 2003 2004 2005 2010 Average observed in the photographs. Conversely, the decline in grass Year cover between 1998 and 2004 was associated with an increase b in dwarf shrub cover. Fluctuations in dwarf shrub cover 100 90 VM04 were less extreme than for grasses which changed by 60 80 % 70 VM05 or more at a site. The response of grass cover appeared more 60 VM13 50 pronounced in Grassy Dwarf Shrubland sites than Dwarf 40 VM14 30 VM22 Shrubland sites. Tall shrub cover was lower in Grassy Dwarf 20 VM24 10 Shrubland at 6% compared to the Dwarf Shrubland sites at 0 VM25 Cover of dwarf shrubs of dwarf Cover 1992 1994 1995 1996 1997 1998 2003 2004 2005 2010 Average

22%, but changed very little temporally for both vegetation % units over the monitoring period. Year c 100 Change in growth form cover within different 90 80 VM04 landforms 70 VM05 60 The three growth forms responded differently between VM13 50 VM14 40 1988 and 2010 within the main landforms at the CNP 30 VM22 (Figure 11a–i). Within rivers, the cover of grasses increased 20 VM24 10 VM25 Cover of tall shrubs of tall Cover 0 from an average of 30% – 50% over the study period. In Average % 1992 1994 1995 1996 1997 1998 2003 2004 2005 2010 contrast, the cover of grasses on the plains increased from Year 30% – 45% between 1994 and 1996, but declined thereafter FIGURE 9: Percent 7: Percentageage covercover perper sitesite cover ((±± s.d.s.d.)) ofof per,, (a)(a) ggrasses, siterasses, (±(b)(b) d d s.d.)warfwarf s s hrubshrubs of, and (a)and (c) (c) grasses, t tallall s shrubshrubs in (b)in Dwarf Dwarf dwarf Shrublands Shrublands shrubs (Nama (Nama - - karoo Biome) as determined from an analysis of fixed-point photo monitoring sites between 1992 and 2010. to less than 20% in 2003, increasing again to reach average FIGUREkaroo 9:Biome) Percent as agedetermined cover per from site an(± s.d.analysis) of, (a)of fixedgrasses,-point (b) photo dwarf monitoring shrubs and sites (c) t allbetween shrubs 1992 in Dwarf and 2010.Shrublands (Nama- and (c) tall shrubs in Grassy Dwarf Shrublands (Grassland Biome) as determined karoofrom Biome) an asanalysis determined of from fixed-point an analysis of fixedphoto-point monitoringphoto monitoring sites sites betweenbetween 1992 1992 and 2010. and 2010. values of 35% in 2010. Whilst grass cover also fluctuated on

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 8 of 16 Original Research

a b c

VM25 N 1994 Match VM25 N 1992 Match VM25 N 1996 Match d e f

VM25 N 1998Source Match: Photograph by Peter Burnett VM25 N 2003 Match VM25 N 2010 Match FIGURE 8: Grassy Dwarf Shrublands example (VM25 N) showing cyclical pattern and switch between grass and dwarf shrub dominance at the Camdeboo National Park, whilst tall shrubs have remained stable from 1992 until 2010. (a) 1992, (b) 1994, (c) 1996, (d) 1998, (e) 2003 and (f) 2010.

a the slopes, responses were far less extreme and remained 100 between 10% and 25%. The cover of dwarf shrubs was 90 VM06 80 VM10 relatively stable on all landforms, with slopes and rivers 70 VM15 60 showing less variation over time than on the plains. Average VM18 50 cover values for this growth form were below 20 for rivers 40 VM21 %

Grass cover Grass 30 VM30 and slopes but around 50% on the plains. Changes in dwarf

% 20 VM31 10 VM32 shrub cover on the plains in particular appeared to respond 0 1992 1994 1995 1996 1998 2003 2004 2005 2010 Average in an opposite direction to changes in grass cover. Tall shrub Year cover dominated the slopes (60%), was co-dominant within rivers (35%) and comprised a relatively minor component of 100 b 90 the cover for plains environments (15%). Tall shrub and tree 80 VM06 70 VM10 cover was relatively stable for plains and slopes over time. 60 VM15 Within rivers, tall shrub cover declined between 1998 and 50 VM18 40 VM21 2003 but increased again over the next 7 years. 30 VM30 20 10 VM31 0 VM32 Change in historical climate Cover of dwarf shrubs of dwarf Cover 1992 1994 1995 1996 1998 2003 2004 2005 2010 Average % Year Annual rainfall No significant change in annual rainfall was recorded over 100 c 90 the last 18 years at the four climate stations within the CNP 80 VM06 (Online Appendix 1; Figure 1). This pattern is consistent 70 VM10 shrubs 60 VM15 with the longer rainfall record for Graaff-Reinet which has

tall 50 VM18 not changed significantly in the last 100 years (Masubelele 40 VM21 30 VM30 2012). Annual rainfall totals varied across the landscape. 20 VM31 They were highest for high elevation sites such as the Valley

Cover of Cover 10 0 VM32 of Desolation (1243 m.a.s.l.) and Waaihoek (1276 m.a.s.l.) (not % 1992 1994 1995 1996 1998 2003 2004 2005 2010 Average included in Figure 12, as records for these sites ceased at the Year end of 2005), which both had values of 358 mm. Sites slightly FIGURE 9: Percentage cover per site (± s.d.) of, (a) grasses, (b) dwarf shrubs and lower in elevation (all below 820 m.a.s.l.), such as Nqweba (c) tall shrubs in Dwarf Shrublands (Nama-karoo Biome) as determined from an analysis of fixed-point photo monitoring sites between 1992 and 2010. Dam (345 mm) and Wildebeeskom (340 mm), received lower

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127

Page 9 of 16 Original Research

a b c

VM18 E 1992 Match d VM18 E 1994 Match e VM18 E 1996 Match f

Source: Photograph by Peter Burnett VM18 E 2003 Match VM18 E 2010 Match VM18FIGURE E 1998 10: Match Dwarf shrublands example (VM18 N) showing a less pronounced cyclic patterns and switches between grasses and dwarf shrubs dominance from 1992 until 2010. (a) 1992, (b) 1994, (c) 1996, (d) 1998, (e) 2003 and (f) 2010.

100 a 100 b 100 c 90 90 90 80 80 80 70 70

70 shrubs 60 60 60 50 50 tall 50 40 40 40 30 shrubs 30 30 Grass cover Grass 20 20 20 Cover of dwarf of dwarf Cover

% 10 10 10 0 % 0 0 Cover of Cover

1992 1994 1995 1996 1998 2003 2004 2005 2010 1992 1994 1995 1996 1998 2003 2004 2005 2010 % 1992 1994 1995 1996 1998 2003 2004 2005 2010 Year Year Year

100 d 100 e 100 f 90 90 90 80 80 80 70 70 70 60 60 shrubs 60 50 50 50 40 40 tall 40 30 30 30 shrubs Grass cover Grass 20 20 20 10 10 10 % 0 of dwarf Cover 0 0 % Cover of Cover 1989 1998 2003 2005 2010 1989 1992 1994 1995 1996 1998 2003 2004 2005 2010 1989 1992 1994 1995 1996 1998 2003 2004 2005 2010

1992 1994 1995 1996 2004 % Year Year Year

100 g 100 h 100 i 90 90 90 80 80 80 70 70 70 shrubs 60 60 60

50 50 tall 50 40 40 40 30 30 30 shrubs

Grass cover Grass 20 20 20 10 Cover of dwarf of dwarf Cover 10 10 % 0 0 0 Cover of Cover %

1989 1992 1994 1995 1996 1998 2003 2004 2005 2010 1989 1992 1994 1995 1996 1998 2003 2004 2005 2010 % 1989 1992 1994 1995 1996 1998 2003 2004 2005 2010 Year Year Year FIGURE 11: Percentage cover (± s.d.) of the main growth forms on the different landforms, rivers (a–c), plains (d–f) and slopes (g–i), as determined from an analysis of fixed-point photo monitoring sites between 1992 and 2010 for rivers and 1989 and 2010 for plains and slopes.

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 10 of 16 Original Research mean annual totals over the course of the study period. Pan evaporation The lowest mean annual rainfall totals were recorded at There was no significant trend between 1926 and 2008 in Gannaleegte (294 mm) and Spekbos (293 mm), which were the long-term S-pan evaporation data for Nqweba Dam. also the lowest in elevation. However, when analysed for the period from 1989 to 2008 only, there was a significant decline in S-pan evaporation Seasonal rainfall values at Nqweba Dam (Online Appendix 1; Figure 3). The Monthly rainfall data for all four weather stations were grouped and the seasonal averages compared between two 3 y = -0.0017x + 0.2224 a time periods (1990–1999 and 2000–2009). Results showed that R2 = 0.0162 p = NS the average amount of rain in summer (December–February) 2 and autumn (March–May) was significantly greater for the 1 decade between 2000 and 2009 than for the decade which preceded it (Figure 12). The increase in both summer and 0 autumn rain was largely the result of the significant increase SPI Index in rain in February and in April, respectively. Spring and -1 winter rainfall amounts were not significantly different -2 between the two decades. -3 Standardised precipitation index y = 0.0003x − 0.0447 b 3 There was no significant change in the long-term SPI at R2 = 0.0007 p = NS all four climate stations (Figure 13a–d). The most severe 2 drought years over the study period were 1990–1991, 1997–1998 1 and 2002–2003. The wettest periods were experienced in

1988, 1995, 2000 and 2006–2007. All four climate stations 0

showed similar SPI profiles, suggesting a high degree of SPI Index spatial coherence in drought and wet periods across the -1 CNP. -2

Temperature -3

Although the average maximum temperature has increased y = -0.0014x + 0.1908 c 3 by 0.001 ºC per year since 1992 and the average minimum R2 = 0.012 p = NS temperature by 0.04 ºC per year, the trends were not 2 significant (Online Appendix 1; Figure 2). The highest monthly maximum temperature was recorded in January 1

2003 at 34.6 ºC and the highest minimum temperature 0 recorded in 2006 at 16.6 ºC. The highest average annual SPI Index maximum temperature of 27.2 ºC was recorded in 1999. The -1 lowest was recorded in 1996 at 25.4 ºC. -2

Dec. – Feb. -3 160 b Mar. – May d 140 June – Aug. 3 y = 0.001x + 0.1362 Sept. – Nov. R2 = 0.0061 a p = NS 120 b 2 a a 100 a 1 80 a 60 a 0 SPI Index 40 -1 20

Average monthly rainfall (mm) rainfall monthly Average -2 0 1990–1999 2000–2009 -3 Time period SPI, standardised precipitation index; NS, not significant. FIGURE 12: Difference in seasonal rainfall per decade at the Camdeboo National Park based on the results from four climate stations in the area. FIGURE 13: Standardised precipitation index values based on a 24-month Dissimilar superscripts denote significant seasonal differences between the period for four stations, (a) Spekbos, (b) Gannaleegte, (c) Wildebeeskom and (d) decades at p < 0.05. Nqweba Dam at the Camdeboo National Park.

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 11 of 16 Original Research maximum annual value for pan evaporation was recorded in Of all the growth forms, grasses were the most dynamic 1993 at 2237 mm, whilst the minimum value was recorded in in terms of how they changed over time as compared to the year 2009 at 1643 mm. the changes in dwarf and tall shrubs. Whilst such changes were somewhat muted in Albany Thicket vegetation, they Animal distribution within the Cambedoo were most noticeable within Azonal vegetation associated National Park with rivers and floodplains, as well as on the plains where Park scale Grassy Dwarf Shrubland and Dwarf Shrubland vegetation Over the course of the study, the highest herbivore stocking dominated. The response of dwarf shrubs in all vegetation densities were recorded around the dam as well as near units appeared different to that of grasses and generally watering points on the plains, whilst mountainous areas declined when grass cover increased. Whilst the cover of had the lowest densities (Online Appendix 1; Figure 4a–u). dwarf shrubs might have been underestimated from the During the period between 1990 and 1994, when rainfall was photographs because of the presence of taller grasses which below average, stocking densities remained high and animals might have obscured their view, the survey data in O’Connor were widely distributed throughout the park. Animals and Roux (1995) suggests that grasses and shrubs compete numbers have declined since 1995 (Figure 14a–d) and they for both above-ground and below-ground resources. Their have also become less widely distributed in the landscape. survey data, as well as evidence presented in Masubelele The plains and the area around the dam support large bulk et al. (in review), show that when grass cover is high, shrub grazers such as Syncerus caffer (buffalo), Damaliscus pygargus cover declines. phillipsi (blesbuck), Connochaetes gnou (black wildebeest) and Taurotragus oryx (eland) which are resident in these areas. Shiponeni et al. (2011) have shown experimentally how Antidorcas marsupialis (springbok) are mixed feeders and are grasses and shrubs compete for resources. Their results also present at high densities in these areas of the park. The indicate that grasses are superior competitors and can out- plains and the area around the dam have a relatively high compete dwarf shrubs in the ecotonal environments that cover of grasses and animal densities have remained high occur between the arid Namaqualand and Bushmanland compared to other areas in the park. In terms of browsers, regions in the western part of the Karoo. They present a Tragelaphus strepsiceros (kudu) is widespread throughout the mechanism and suggest that increased grass cover results park as well as within the Nqweba Dam area where it feeds in reduced soil water availability for the nearest shrubs. on A. karroo. Tall shrubs, such as P. afra which are dominant Alvarez et al. (2011) have also shown within the Chihuahuan within the more mountainous areas of the CNP, are also desert on the US–Mexican border that when rainfall is heavily utilised by kudu. abundant, increased competition occurs between shrubs and grasses. The cover assessments derived from a comparison Herbivore densities associated with vegetation units of repeat photographs in this study thus appear to reflect Average herbivore densities within the vegetation units real differences on the ground between shrub and grass showed that Azonal vegetation and the associated Grassy cover (Hoffman & Rohde 2011a). However, more detailed Dwarf Shrubland areas supported the highest densities of experimental work is needed to determine the competitive herbivores, whereas the Dwarf Shrublands and the Albany interactions between grasses and shrubs within eastern Karoo Thicket had the lowest (Figure 14a–d). Sites within Azonal rangelands, where fluctuations in grass and shrub cover can vegetation had higher densities during the drought years of occur over relatively short time periods (Hoffman et al. 1990). 1990–1992, 1997–1998 and 2003, whereas 1993 and 2004 had Experimentally induced drought and above-average rainfall the highest densities of all. The surrounding Grassy Dwarf events would be especially helpful in this regard and would Shrublands had higher stocking densities immediately serve to test the state-and-transition model proposed by after the drought years. The Dwarf Shrublands and Albany Milton and Hoffman (1994) for the region. Thicket communities, on the other hand, have supported relatively low densities of animals over the last 20 years. The cover of tall shrubs in all vegetation types except Azonal vegetation was remarkably stable over the course of the Discussion study. With few exceptions, the cover of tall shrubs in 2010, at almost every site within Albany Thicket, Grassy Dwarf Change in vegetation cover and growth form Shrublands and Dwarf Shrublands, was the same or very composition since 1992 similar to what it was at the beginning of the study when Vegetation cover and growth form composition are the most the first photographs were taken. This was not the case, commonly used indicators of change in most terrestrial however, for several sites within Azonal vegetation, where ecosystems (Godinez-Alvarez et al. 2009). In this study, there tall shrub cover responded relatively quickly to fluctuations was considerable variation in the total cover of vegetation, in the hydrological environment. Acacia karroo appeared as well as the cover of different growth forms between sites particularly dynamic on the flood plain above the Nqweba and even within the same vegetation unit. Although this Dam. For example, the dry period between 1998 and 2003 variability makes it difficult to generalise across the CNP, had a negative impact on the abundance of A. karroo, which some common patterns were evident in the response of appeared unable to survive long periods of hydrological different growth forms within different vegetation units and drought. Other studies report a similar decline in tree and landforms. shrub cover under drought conditions (Allen et al. 2010;

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 AOSISAOSIS OpenJournals OpenJournals – Language – Language Edited Edited Version Version 14 14Aug Aug 2013 2013

AOSIS OpenJournals – Language Edited Version 14 Aug 2013 AOSIS OpenJournals – Language Edited Version 14 Aug 2013 FigureFigure 14 14 Page 12 of 16 AOSISOriginal OpenJournals Research – Language Edited Version 14 Aug 2013 Figure 14 10

Figure) 14 a Figure2 14 VM01 8 VM02 VM03 VM23 6 VM26 AOSIS OpenJournals – Language Edited VersionVM27 14 Aug 2013 VM28 Average 4 Figure 14 2 AOSIS OpenJournals – Language Edited Version 14 Aug 2013 AOSIS OpenJournals – Language Edited Version 14 Aug 2013

Average herbivore density (LSU/km herbivore Average 0 1990 1991 1992 1993 1994 1995 1996 1997 1998Figure 1999 2000 14 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 FigureYears 14 20 b ) 2 VM04 VM05 VM13 15 VM14 VM22 VM24 VM25 10 Average

Average herbivore density (LSU/km herbivore Average 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Years

10 c ) 2 VM06 VM10 8 VM15 VM18 VM21 6 VM30 VM31 VM32 4

Average herbivore density (LSU/km herbivore Average 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Years 10 d ) 2 VM07 8 VM08 VM09 VM11 VM12 6 VM16 VM17 VM19 VM20 4 VM29 Average

Average herbivore density (LSU/km herbivore Average 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Years LSU, large stock equivalent. FIGURE 14: Average herbivore stocking density per site (LSU/km²) from 1990 until 2010 recorded at the 32 vegetation monitoring sites in the Camdeboo National Park. Note the different y-axis scale for Grassy Dwarf Shrublands, with (a) representing Azonal vegetation, (b) Grassy Dwarf Shrublands, (c) Dwarf Shrublands and (d) Albany Thicket.

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 13 of 16 Original Research

Fensham, Fairfax & Ward 2009). In the CNP, however, The two flood events that occurred in 2001 and 2006, for populations of A. karroo responded quickly again to the example, resulted in an increase in grasses in subsequent increase in moisture levels after the relatively good rains years. The floods were also probably responsible for the experienced in the region in the 2007–2008 rainy season. cohort of smaller A. karroo and T. rammosissima saplings Hoffman and Rohde (2011b) have indicated the dynamic which became evident soon after these events. Drought nature of river systems in Namaqualand which have also events, such as were recorded in 1998 and 2003, also shown a significant increase in the abundance of A. karroo influence vegetation dynamics and, in the CNP, contributed over time. However, their repeat photograph study, which to a decline in grass cover but not that of dwarf shrubs. This usually only has two time steps, does not capture the dynamic was particularly evident in the Grass Dwarf Shrublands nature of river systems in arid environments where shorter and Dwarf Shrublands, where grass–shrub interactions are observation periods of 5 or 10 year time steps are preferable. mediated in part by extreme drought events (Hoffman & Whilst riparian vegetation communities are generally Cowling 1990; Snyder & Tartowski 2006). acknowledged to be spatially and temporally dynamic as a result of fluvial disturbance amongst other drivers (Perryet al. Evidence from the photo-monitoring assessment also 2012), there is a lack of long-term monitoring data to show suggests that in arid and semi-arid environments drought the nature, extent and rate of this dynamism (Webb et al. 2011). years can significantly influence the structure of riparian and azonal habitats by eliminating large trees, especially when Changes in climate and its influence on the droughts occur in combination with increased temperature vegetation (Allen et al. 2010; Breshears et al. 2009). In this study, Rainfall drought and higher temperatures in 1998 and 2003 might Trends in climate parameters measured at the CNP over the have influenced the structure of the vegetation within the last 20 years are consistent with those reported elsewhere Nqweba Dam floodplain through the elimination of older for southern Africa (Masubelele 2012; Warburton, Schulze & and therefore taller individuals of A. karroo. Maharaj 2005). For example, there was no significant trend in annual rainfall or the incidence of drought at any of the Temperature and pan evaporation climate stations within the CNP over the course of the study As is the case in many parts of southern Africa over the period. There were, however, significant fluctuations in last 20 years (Engelbrecht, McGregor & Engelbrecht 2009; annual rainfall between years, with important implications Warburton et al. 2005), minimum temperature has increased at for the vegetation of the region (Milton & Hoffman 1994). the CNP at a faster rate than maximum temperature. Despite The analysis of seasonal rainfall patterns suggests that the this, however, pan evaporation values declined significantly amount of rain falling in late summer and autumn increased from 1989. This decline has been reported elsewhere in in the second decade of the study when compared to the southern Africa (Eamus & Palmer 2007; Hoffman et al. 2011) first, although the high degree of inter-annual variability and appears part of a global phenomenon (Roderick, Hobbins might have had a significant influence on this pattern. This & Farquhar 2009) which has, as yet, not been adequately appeared not to be the case for spring and winter rainfall explained (McVicar et al. 2012). In the winter rainfall region patterns which did not differ significantly over the course of South Africa, a significant decline in wind run appears the of the study. This finding is consistent with Du Toit (2010), most plausible explanation (Hoffman et al. 2011) but more who reported a similar increase in summer rain at a climate work is needed to resolve this issue. station at Middleburg about 100 km north of the CNP. Herbivory and its impact on the vegetation Inter-annual fluctuations in rainfall have important implications for the response of different growth forms in An understanding of the spatial and temporal use of the the region as evidence by the repeat photograph survey. landscape by the herbivores is an integral part of ecosystem Grasses tend to respond to short-term fluctuations in annual management (Bailey et al. 1996). Herbivores are known to rainfall far more than tall shrubs and trees (Holdo, Holt & affect ecosystem structure through both the loss and increase Fryxell 2009; Miranda et al. 2009; Scanlon et al. 2005; Snyder of tree cover and to influence ecosystem processes such as & Tartowski 2006). Summer rainfall, derived from convective productivity and carbon storage at local to regional scales thunderstorms, often contributes short-term pulses to soil (Tanentzap & Coomes 2012). It is therefore important to understand the temporal and spatial dynamics of herbivory moisture that can be utilised quickly by annual plants and C4 grasses in particular (Schwinning et al. 2008). Evidence from at community to landscape scale at the CNP. Animal census the photo-monitoring programme suggests that the increase data indicated that ungulate densities have declined over in summer rainfall over the first decade of the 21st century time and have also varied considerably over space within has influenced grass cover within the Grassy Shrubland and the CNP. Whilst herbivore density appears significantly Azonal vegetation units at the CNP, as well as within the influenced by the location of the Nqweba Dam, as well as the dominant landforms with which they are associated. water points within the park, the occurrence of floods and droughts also affects the movement and densities of animals. Extreme rainfall events such as droughts and floods or Herbivore density remained relatively low on the slopes above-normal rainfall influence the temporal dynamics of of the CNP, whilst higher animal densities were generally vegetation as observed in the photo monitoring assessments. recorded on the plains environments nearest to the watering

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 14 of 16 Original Research points. The years with the highest herbivore density included regularly updated and properly archived set of fixed-point 1992, 1995, 1998 and 2004. The majority of the high animal photographs provides a relatively inexpensive and easily density records were returned during drought years. The implemented monitoring. This monitoring technique carried lack of available moisture in the landscape might have forced out over 18 years at 32 locations in the CNP was able to animals to concentrate around artificial water points in the document short-term changes in growth forms in different park and would have influenced the ease with which animals vegetation units and land forms and long-term fluctuations were observed and counted. of several individual species as well as alien plants. These data, when collected over decades, are essential in separating Animal densities were highest in the Azonal vegetation directional change from short-term variability in the followed by the Grassy Dwarf Shrublands. The location of environment. Coupled with detailed rainfall and livestock several artificial water points in this vegetation unit, as well as census records, the relative importance of the underlying the good cover of high quality grasses, could have influenced drivers of change can also be determined. This is the sort of this. It is well known that both slope and distance to water information that park managers consider necessary for both points can exert a major influence on the movement and their day-to-day management decisions and their long-term distribution of herbivores (Bailey et al. 1996). The density of strategic planning activities. A sound historical assessment of herbivores was significantly lower on slopes, with only areas the pattern of change can also help to assess possible future near water points having slightly higher herbivore densities. trajectories based on land use and climate change scenarios This is largely a result of the difficulty and restriction of (Hendrick & Copenheaver 2009; Munson, Webb & Hubbard movement by the dense vegetation on steep slopes, which 2011; Peters et al. 2011; Webb et al. 2011). also reduces the ability of rangers to observe and record animals during the census activities (Morellet et al. 2007). However, the use of repeat photographs within a long-term However, even within vegetation units and landforms monitoring programme is not without its problems. For which supported relatively high densities of animals, there example, whilst the estimation of vegetation cover from is little evidence from the photo-monitoring assessment that photographs can be accurate and similar in precision with herbivory played a significant role in influencing the cover point-frequency estimates (Symstad, Wienk & Thorstenson of the dominant growth forms or species. Fluctuations in 2008), it is influenced by observer bias. Experienced observers tend to underestimate vegetation cover, whilst rainfall and major changes in the hydrological dynamics of less-experienced observers often overestimate this value the Nqweba Dam floodplain appeared far more important (Lucey & Barraclough 2001). It is also often difficult to identify in influencing the cover of grasses, dwarf shrubs and tall smaller shrubs in the landscape when grass cover is high, shrubs than changes in animal densities. However, the use which leads to an underestimation of this growth form in the of a photo-monitoring approach might be an inappropriate photograph, particularly in the Grassy Dwarf Shrublands in tool to measure the impact of herbivory on the structure of this article. Similarly, large trees, especially when they have the CNP ecosystems. As for other semi-arid conservation established close to the camera station, can also obscure the areas (Hoffman et al. 2009), it appears that herbivore densities field of view as was the case for some images within the currently observed at the CNP are too low to have had a Azonal and Albany Thicket vegetation units photographed major influence on the vegetation, especially when compared in the CNP. to those commonly recorded for livestock in surrounding areas. The selective feeding nature of wild herbivore guilds, This study has also emphasised that a set of repeat compared to the concentrated heavy grazing and browsing photographs without more detailed survey data is of limited associated with livestock (Laca et al. 2010), may explain why value only. The inclusion of vegetation survey data at each the impact of animals on the vegetation of the CNP appears camera station along the direction of the field of view adds to have been negligible (Groom & Harris 2010). However, a further depth to the analysis particularly in terms of an more detailed assessment of the spatial use of the CNP by assessment of changes in species diversity, range condition different herbivore guilds is urgently needed. and other species-dependent metrics of ecosystem structure and function (Lucey & Barraclough 2001). Furthermore, The use of photo monitoring as a management without the vegetation surveys undertaken at each location tool in 2010, little insight into the looming alien plant problem As mentioned in the ‘Introduction’, there is a need to initiate (Masubelele et al. 2009), nor the efficacy of the alien clearing and support biodiversity monitoring frameworks in South programme would have been gained. Associated metadata, Africa and abroad (Scholes et al. 2008; Reyers & McGeoch such as stocking rates, rainfall and specific management 2007). The photo-monitoring data from this study should interventions, also help in the interpretation of landscape- contribute not only to SAEON data requests at a national level changes and are essential components for any photo- scale but also to our protected areas aspirations towards monitoring programme. The nature of these monitoring achieving their monitoring frameworks (McGeoch et al. tools, exactly what is recorded, as well as the frequency of 2011). Park managers use photos regularly to inform their measurement will probably differ between conservation decisions and to effect policy change (Hall 2000; Hendrick areas. However, monitoring approaches could be similar & Copenheaver 2009; Lucey & Barraclough 2001; Nichols within the programmes developed for groups of national et al. 2009; Turner et al. 2003). This study has shown that a parks such as those in the arid and semi-arid areas of South

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 15 of 16 Original Research

Africa. Such a standardised programme will not only National Park. M.L.M. analysed the animal census data in contribute to SANParks monitoring efforts, but will also both ArcGIS and Excel. M.T.H. and W.B. (University of Cape heed the call for the development of national monitoring Town) made conceptual contributions to the project design programmes (Scholes et al. 2008). and also commented on the draft manuscript, which was written by M.L.M. as part of his PhD thesis. Conclusion References Using a repeat photo-monitoring approach this study has documented the varied response of vegetation cover within Allen, C.D., Macalady, A., Chenchouni, H., Bachelet, D., McDowell, N., Vennetier, M.et al., 2010, ‘A global overview of drought and heat-induced tree mortality reveals three main growth forms in different major vegetation units emerging climate change risks for forests’, Forest Ecology & Management 259, 660–684. http://dx.doi.org/10.1016/j.foreco.2009.09.001 and landforms within the CNP since 1988. Sites within Alvarez, L., Epstein, H.E., Li, J. & Okin, G.S., 2011, ‘Spatial patterns of grasses and Albany Thicket and Dwarf Shrublands showed the least shrubs in an arid grassland environment’, Ecosphere 2, art103. change in vegetation cover, whilst Azonal vegetation and ArcGIS version 9.3, 2008, computer software, Esri, Redlands. Grassy Dwarf Shrublands were more dynamic in their Bailey, D.W., Gross, J.E., Laca, E.A., Rittenhouse, L.R., Coughenour, M.B., Swift, D.M.et al., 1996, ‘Mechanisms that result in large herbivore grazing distribution patterns’, response to rainfall and flood plain hydrology. Abiotic Journal of Range Management 49, 386–400. http://dx.doi.org/10.2307/4002919 factors such as rainfall amount, changes in the distribution Baraza, E., Zamora, R. & Hódar, J.A., 2006, ‘Conditional outcomes in plant-herbivore interactions: Neighbours matter’,Oikos 113, 148–156. http://dx.doi.org/10.1111/ of rain between seasons and the occurrence of flood and j.0030-1299.2006.14265.x drought events in combination with high temperatures Breshears, D.D., Myers, O.B., Meyer, C.W., Barnes, F.J., Zou, C.B., Allen, C.D. et al., 2009, ‘Tree die-off in response to global change-type drought: Mortality insights appeared to have the greatest influence on growth form from a decade of plant water potential measurements’, Frontiers in Ecology and dynamics. Grasses generally responded to above-average the Environment 7, 185–189. http://dx.doi.org/10.1890/080016 rainfall events and because of their superior competitive Bruner, A.G., Gullison, R.E., Rice, R.E. & Da Fonseca, G.A.B., 2001, ‘Effectiveness of parks in protecting tropical biodiversity’, Science 291, 125−128. http://dx.doi. ability, negatively influenced the cover of dwarf shrubs. org/10.1126/science.291.5501.125, PMid:11141563 Dwarf shrubs appeared less affected by drought conditions, Burdett, P.D., 1995, Karoo Nature Reserve management plan, unpublished report, Eastern Cape Nature Conservation, Bisho. whilst grass cover declined. The cover of tall shrubs and Camdeboo National Park Management Plan, 2006, The management plan of trees declined under drought conditions and increased Camdeboo National Park, unpublished report, South African National Parks, significantly after major flood events, particularly within the Graaff-Reinet. Camdeboo National Park Newsletter, 2011, The Camdeboo National Park news, Issue flood plain associated with the Nqweba Dam. Biotic factors 2, July, p. 4, unpublished report, South African National Parks, Graaff-Reinet. such as herbivory appear to have had a negligible impact Department of Environmental Affairs, 2010, South Africa’s second national communication under the United Nations Framework Convention on Climate on the vegetation of the CNP, at least in terms of how this Change, Department of Environmental Affairs, Pretoria. could be determined from the photo-monitoring techniques Driver, A., Sink, K.J., Nel, J.N., Holness, S., Van Niekerk, L., Daniels, F. et al., 2012, National biodiversity assessment 2011: An assessment of South Africa’s used in this study. The lack of any visible impact is primarily biodiversity and ecosystems, synthesis report, South African National Biodiversity because animal densities have remained low over the course Institute and Department of Environmental Affairs, Pretoria. of the study, relative to values for commercial livestock Du Toit, J.C.O., 2010, ‘An analysis of long-term daily rainfall data from Grootfontein, 1916 to 2008’, Grootfontein Agric 10, 24–36. production systems in the area. Finally, the potential for using Eamus, D. & Palmer, A.R., 2007, ‘Is climate change a possible explanation for woody a photo-monitoring programme to document the nature, thickening in arid and semi-arid regions?’, Research Letters in Ecology 2007, 1–5. http://dx.doi.org/10.1155/2007/37364 extent and rate of vegetation change over decades is clearly Ehrlich, P.R. & Pringle, R.M., 2008, ‘Where does biodiversity go from here? A demonstrated. However, the value of a well maintained, grim business-as-usual forecast and a hopeful portfolio of partial solutions’, Proceedings of the National Academy of Sciences USA 105, 11579–11586. http:// properly archived and easily accessible collection of repeat dx.doi.org/10.1073/pnas.0801911105, PMid:18695214, PMCid:PMC2556413 photos would be considerably enhanced by the collection Engelbrecht, F.A., Mcgregor, J.L. & Engelbrecht, C.J., 2009, ‘Dynamics of the conformal- cubic atmospheric model projected climate-change signal over southern Africa’, of additional field survey data at the same time that the International Journal of Climatology 29, 1013–1033. http://dx.doi.org/10.1002/ photographs were taken. Well-designed experiments are joc.1742 also needed to determine the relative influence of different Fensham, R.J., Fairfax, R.J. & Ward, D.P., 2009, ‘Drought-induced tree death in savanna’, Global Change Biology 5, 380–387. http://dx.doi.org/10.1111/j.1365- drivers identified in this study at community and landscape 2486.2008.01718.x scales. Furstenburg, D., 2002, ‘Optimizing game production in a new era’, Karoo Agric 5, 17–27. Gitzen, R.A., Millspaugh, J.J., Cooper, A.B. & Licht, D.S. (eds.), 2012, Design and analysis of long-term ecological monitoring studies, Cambridge University Press, Acknowledgements Cambridge. http://dx.doi.org/10.1017/CBO9781139022422 Godínez-Alvarez, H., Herrick, J.E., Mattocks, M., Toledo, D. & Van Zee, J., 2009, Competing interests ‘Comparison of three vegetation monitoring methods: Their relative utility for ecological assessment and monitoring’, Ecological Indicators 9, 1001–1008. The authors declare that they have no financial or personal http://dx.doi.org/10.1016/j.ecolind.2008.11.011 relationships that may have inappropriately influenced them Groom, R. & Harris, S., 2010, ‘Factors affecting the distribution patterns of zebra and wildebeest in a resource-stressed environment’, African Journal of Ecology 48, in writing this article. 159–168. http://dx.doi.org/10.1111/j.1365-2028.2009.01097.x Hall, F.C., 2000, Photo point monitoring handbook, General technical reports of the US Department of Agriculture, Forest Service, Washington, DC. Authors’ contributions Hendrick, L. & Copenheaver, C., 2009, ‘Using repeat landscape photography to assess vegetation changes in rural communities of the southern Appalachian Mountains M.L.M. (University of Cape Town) completed the in Virginia, USA’, Mountain Research & Development 29, 21–29. http://dx.doi. photography in 2010, with initial assistance from M.T.H. org/10.1659/mrd.1028 Hoffman, M.T. & Cowling, R.M., 1990, ‘Desertification in the lower Sundays River (University of Cape Town), and also performed the analysis Valley, South Africa’, Journal of Arid Environments 19, 105–117. of climate data, whereas P.B. (Camdeboo National Park) set Hoffman, M.T. & Rohde, R.F., 2011a, ‘Long-term changes in the vegetation of southern up the sites and photographed the locations from 1988 to 2005. Africa as revealed by repeat photography’, in L. Zeitsman (ed.), Observations of large-scale and long-term trends in southern Africa, pp. 79–83, South African P.B. also provided most of the datasets for the Camdeboo Environmental Observation Network, Pretoria.

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127 Page 16 of 16 Original Research

Hoffman, M.T. & Rohde, R.F., 2011b, ‘Rivers through time: Historical changes in the Munson, S.M., Webb, R.H. & Hubbard, J.A., 2011, ‘A comparison of methods to assess riparian vegetation of the semi-arid, winter rainfall region of South Africa in long-term changes in Sonoran desert vegetation’,Journal of Arid Environments 75, response to climate and land use’, Journal of the History of Biology 44, 59–80. 1228–1231. http://dx.doi.org/10.1016/j.jaridenv.2011.04.032 http://dx.doi.org/10.1007/s10739-010-9246-4 Murwira, A. & Skidmore, A.K., 2005, ‘The response of elephants to the spatial Hoffman, M.T., Barr, G.D. & Cowling, R.M., 1990, ‘Vegetation dynamics in the semiarid heterogeneity of vegetation in a Southern African agricultural landscape’, eastern Karoo, South Africa – The effect of seasonal rainfall and competition on Landscape Ecology 20, 217–234. http://dx.doi.org/10.1007/s10980-004-3159-6 grass and shrub basal cover’, South African Journal of Science 86, 462–463. Nichols, M.H., Ruyle, G.B. & Nourbakhsh, I.R., 2009, ‘Very high resolution panoramic Hoffman, M.T., Cramer, M.D., Gillson, L. & Wallace, M., 2011, ‘Pan evaporation photography to improve conventional rangeland monitoring’, Rangeland Ecology and wind run decline in the Cape Floristic Region of South Africa (1974–2005): & Management 62, 579–582. http://dx.doi.org/10.2111/.1/REM-D-09-00017.1 Implications for vegetation responses to –climate change’, Climatic Change 109, Noss, R.F., 1990, ‘Indicators for monitoring biodiversity: A hierarchical approach’, 437–452. http://dx.doi.org/10.1007/s10584-011-0030-z Conservation Biology 4, 355–364. http://dx.doi.org/10.1111/j.1523-1739.1990. Hoffman, M.T., Madden, C.F., Erasmus, K., Saayman, N. & Botha, J.C., 2009, ‘The tb00309.x impact of indigenous ungulate herbivory over five years (2004–2008) on the O’Connor, T.G. & Roux, P.W., 1995, ‘Vegetation changes (1949–1971) in a semi-arid, vegetation of the Little Karoo, South Africa’, African Journal of Range and Forage Grassy Dwarf Shrubland in the Karoo, South Africa: Influence of rainfall variability Science 26, 169–179. http://dx.doi.org/10.2989/AJRF.2009.26.3.8.953 and grazing by sheep’, Journal of Applied Ecology 32, 612–626. http://dx.doi. Holdo, R.M., Holt, R.D. & Fryxell, J.M., 2009, ‘Grazers, browsers, and fire influence the org/10.2307/2404657 extent and spatial pattern of tree cover in the Serengeti’, Ecological Applications Palmer, A.R., 1990, ‘A qualitative model of vegetation history in the eastern Cape 19, 95–109. http://dx.doi.org/10.1890/07-1954.1, PMid:19323175 midlands, South Africa’, Journal of Biogeography 17, 35–46. http://dx.doi. Kraaij, T. & Milton, S.J., 2006, ‘Vegetation changes (1995–2004) in semi-arid Karoo org/10.2307/2845186 shrubland, South Africa: Effects of rainfall, wild herbivores and change in land Perrings, C., Duraiappah, A., Larigauderie, A. & Mooney, H., 2011, ‘The biodiversity use’, Journal of Arid Environments 64, 174–192. http://dx.doi.org/10.1016/j. and ecosystem services science-policy interface’, Science 331, 1139–1140. http:// jaridenv.2005.04.009 dx.doi.org/10.1126/science.1202400, PMid:21330488 Kruskal, J.B., 1964, ‘Nonmetric multidimensional scaling: A numerical method’, Perry, L.G., Andersen, D.C., Reynolds, L.V., Nelson, S.M. & Shafroth, P.B., 2012, Psychometrika 29, 115–129. http://dx.doi.org/10.1007/BF02289694 ‘Vulnerability of riparian ecosystems to elevated CO2 and climate change in arid Laca, E.A., Sokolow, S., Galli, J.R. & Cangiano, C.A., 2010, ‘Allometry and spatial scales and semiarid western North America’, Global Change Biology 18, 821–842. http:// of foraging in mammalian herbivores’, Ecology Letters 13, 311–320. http://dx.doi. dx.doi.org/10.1111/j.1365-2486.2011.02588.x org/10.1111/j.1461-0248.2009.01423.x, PMid:20100240 Peters, D.P.C., Yao, J., Sala, O.E. & Anderson, J., 2011, ‘Directional climate change Lindenmayer, D.B. & Likens, G.E., 2010, Effective ecological monitoring, CSIRO and potential reversal of desertification in arid and semiarid ecosystems’, Publishing, Collingwood. Global Change Biology 18, 151–163. http://dx.doi.org/10.1111/j.1365- 2486.2011.02498.x The living deserts of southern Africa Lovegrove, B., 1993, , Fernwood Press, Cape Town. Reyers, B. & McGeoch, M.A., 2007, ‘A biodiversity monitoring framework for South Lucey, W.P. & Barraclough, C.L., 2001, ‘A user guide to photopoint monitoring Africa: Progress and directions’, South African Journal of Science 103, 295–300. techniques for riparian areas. Field test edition’, Aqua-Tex Consulting Ltd, Robertson, M.P. & Palmer, A.R., 2002, ‘Predicting the extent of succulent thicket under Kimberely. current and future climate scenarios’, African Journal Range & Forage Science 19, Maiorano, L., Falcucci, A. & Boitani, L., 2008, ‘Size-dependent resistance of protected 21–28. http://dx.doi.org/10.2989/10220110209485771 areas to land-use change’, Philosophical Transactions of the Royal Society of Roderick, M.L., Hobbins, M.T. & Farquhar, G.D., 2009, ‘Pan evaporation trends and the Biological Sciences 275, 1297–304. terrestrial water balance. I. Principles and observations’, Geography Compass 3, Masubelele, M.L., 2012, ‘Understanding the past to conserve the future: Long-term 746–760. http://dx.doi.org/10.1111/j.1749-8198.2008.00213.x environmental and vegetation change in the Karoo Midlands, South Africa over Rodrigues, A.S.L., Andelman, S.J., Bakarr, M.I., Boitani, L., Brooks, T.M., Cowling, R.M. the 20th Century’, PhD Thesis, University of Cape Town, Botany Department, Cape et al., 2004, ‘Effectiveness of the global protected area network in representing Town. species diversity’, Nature 428, 640–643. http://dx.doi.org/10.1038/nature02422, Masubelele, M.L., Foxcroft, L.C. & Milton, S.J., 2009, ‘Alien plant species list and PMid:15071592 distribution for Camdeboo National Park, Eastern Cape Province, South Africa’, Rutherford, M.C., Powrie, L.W. & Husted, L.B., 2012a, ‘Plant diversity consequences of Koedoe 51(1), Art. #515, 10 pages. a herbivore-driven biome switch from grassland to Nama-Karoo shrub steppe in Masubelele, M.L., Hoffman, M.T., Bond, W.J. & Gambiza, J., in review, ‘A 50 year study South Africa’, Applied Vegetation Science 15, 14–25. http://dx.doi.org/10.1111/ shows grass cover has increased in shrublands in semi-arid South Africa’, Journal j.1654-109X.2011.01160.x of Arid Environments. Rutherford, M.C., Powrie, L.W. & Husted, L.B., 2012b, ‘Herbivore-driven land Mather, P.M., 1976, Computational methods of multivariate analysis in physical degradation: Consequences for plant diversity and soils in arid Subtropical Thicket geography, Wiley & Sons, London. PMCid:PMC1639298 in South-Eastern Africa’, Land Degradation & Development 15, 14–25. McCune, B. & Mefford, M.J., 2006, PC-ORD: Multivariate analysis of ecological data. Scanlon, T.M., Caylor, K.K., Manfreda, S., Levin, S.A. & Rodriguez-Iturbe, I., 2005, Version 5.14, MJM Software, Gleneden Beach. ‘Dynamic response of grass cover to rainfall variability: Implications for the function and persistence of savanna ecosystems’, Advances in Water Resources McGeoch, M.A., Dopolo, M., Novellie, P., Hendriks, H., Freitag, S., Ferreira, S. et al., 28, 291–302. http://dx.doi.org/10.1016/j.advwatres.2004.10.014 2011, ‘A strategic framework for biodiversity monitoring in South African National et al. Parks’, Koedoe 53, 1–10. Scholes, R.J., Mace, G.M., Turner, W., Geller, G.N., Jurgens, N., Larigauderie, A. , 2008, ‘Toward a global biodiversity observing system’, Science 321, 1044–1045. McKee, T.B.N., Doesken, J. & Kleist, J., 1993, ‘The relationship of drought frequency http://dx.doi.org/10.1126/science.1162055, PMid:18719268 and duration to time scales’, Proceedings of the 8th Conference On Applied Climatology Schwinning, S., Belnap, J., Bowling, D.R. & Ehleringer, J.R., 2008, ‘Sensitivity of the , American Meteorological Society, Anaheim, 17–22 January, pp. Colorado Plateau to change: Climate, ecosystems, and society’, Ecology & Society 179–184. 13, art.28. McVicar, T.R., Roderick, M., Donohue, R., Li, T.L., Van Niel, T.G., Thomas, A. et al., 2012, Shiponeni, N.N., Allsopp, N., Carrick, P.J. & Hoffman, M.T., 2011, ‘Competitive ‘Global review and synthesis of trends in observed terrestrial near-surface wind interactions between grass and succulent shrubs at the ecotone between an arid speed: Implications for evaporation’, Journal of Hydrology 416–417, 182–205. grassland and succulent shrubland in the Karoo’, Plant Ecology 212, 795–808. http://dx.doi.org/10.1016/j.jhydrol.2011.10.024 http://dx.doi.org/10.1007/s11258-010-9864-0 Meissner, H.H., Hofmeyr, H.S., Van Rensburg, W.J.J. & Pienaar, J.P., 1983, Klassifikasie Snyder, K.A. & Tartowski, S.L., 2006, ‘Multi-scale temporal variation in water availability: van vee vir sinvolle beraming van vervangingswaardes in terme van ‘n biologies- Implications for vegetation dynamics in arid and semiarid ecosystems’, Journal of gedefinieerde Grootvee-eenheid, Tegniese Mededeling, Departement van Arid Environments 65, 219–234. http://dx.doi.org/10.1016/j.jaridenv.2005.06.023 Landbou, RSA, No. 175. Symstad, A.J., Wienk, C.L. & Thorstenson, A.D., 2008, ‘Precision, repeatability, and Midgley, G.F., Hannah, L., Millar, D., Rutherford, M.C. & Powrie, L.W., 2002, ‘Assessing efficiency of two canopy cover estimate methods in northern Great Plains the vulnerability of species richness to anthropogenic climate change in a vegetation’, Rangeland Ecology & Management 61, 413–429. http://dx.doi. biodiversity hotspot’, Global Ecology & Biogeography 11, 445–451. http://dx.doi. org/10.2111/08-010.1 org/10.1046/j.1466-822X.2002.00307.x Tanentzap, A.J. & Coomes, D.A., 2012, ‘Carbon storage in terrestrial ecosystems: Do Milton, S.J. & Hoffman, M.T., 1994, ‘The application of state-and-transition models browsing and grazing herbivores matter?’, Biological Reviews 87, 72–94. http:// to rangeland research and management in arid succulent and semi-arid grassy dx.doi.org/10.1111/j.1469-185X.2011.00185.x, PMid:21635684 Karoo, South Africa’, African Journal of Range & Forage Science 11, 18–26. http:// dx.doi.org/10.1080/10220119.1994.9638349 Trollope, W.S.W., Trollope, L.A., Biggs, H.C., Pienaar, D. & Potgieter, A.L.F., 1998, ‘Long- term changes in the woody vegetation of the Kruger National Park, with special Miranda, J.d.D., Padilla, F.M., Lázaro, R. & Pugnaire, F.I., 2009, ‘Do changes in rainfall reference to the effects of elephants and fire’,Koedoe 41, 103–112. http://dx.doi. patterns affect semiarid annual plant communities?’, Journal of Vegetation org/10.4102/koedoe.v41i2.255 Science 20, 269–276. http://dx.doi.org/10.1111/j.1654-1103.2009.05680.x Turner, R.M., Webb, R.H., Bowers, J.E. & Hastings, J.R., 2003, The changing mile Modarres, R. & Da Silva, V.P.R., 2007, ‘Rainfall trends in arid and semi-arid regions revisited, University of Arizona Press, Tucson. of Iran’, Journal of Arid Environments 70, 344–355. http://dx.doi.org/10.1016/j. jaridenv.2006.12.024 Warburton, M., Schulze, R.E. & Maharaj, M., 2005, ‘Is South Africa’s temperature changing? An analysis of trends from daily records, 1950–2000’, in R.E. Schulze Morellet, N., Gaillard, J.M., Hewison, A.J.M., Ballon, P., Boscardin, Y., Duncan, P. et (ed.), Climate change and water resources in southern Africa: Studies on al., 2007, ‘Indicators of ecological change: New tools for managing populations scenarios, impacts, vulnerabilities and adaptation, pp. 275–295, Water Research of large herbivores’, Journal of Applied Ecology 44, 634–643. http://dx.doi. Commission, Pretoria. org/10.1111/j.1365-2664.2007.01307.x Webb, R.H., Belnap, J., Scott, M.L. & Esque, T.C., 2011, ‘Long-term change in perennial Mucina, L. & Rutherford, M.C. (eds.), 2006, ‘The vegetation of South Africa, Lesotho vegetation along the Colorado River in Grand Canyon National Park (1889–2010)’, and Swaziland’, Strelitzia 19, South African National Biodiversity Institute, Pretoria. Park Science 28(2), 83–87.

http://www.koedoe.co.za doi:10.4102/koedoe.v55i1.1127